Jovian satellites

[Xilman]

Imaging the Galilean satellites is trivial. Where things start getting interesting is when the satellites are either close in (e.g. Amalthea and Thebe) or very faint.

In an effort to prompt other observers to give them I a try this coming season, here is an image of Themisto, aka Jupiter-XVIII which I took during the last apparition. Themisto was discovered in 1975 and promptly lost until 2000. It is tiny, only 9km across, and fainter than 20th magnitude --- 20.3 in this instance. The satellite is marked with cross-bars. The star indicated by an arrow is Gaia EDR3-6771111466513671680 and catalogued at g=20.66.

Exposure: 1708s in 31 subs. Equipment: unfiltered SX814 on a 0.4m Dilworth.

ff

Edited June 24, 2021 by Xilman
Fix s/Themis/Themisto/

[Xilman]

Some other satellites may be forthcoming, either from last year, or from later this year.

[Pete Presland]

On 16/06/2021 at 16:34, Xilman said:

Imaging the Galilean satellites is trivial.

Incredible something so small can be captured from Earth, at such a distance.

[MarkRadice]

Gosh what an amazing image. Would you mind describing your approach to planning and capturing such an observation? It sounds fascinating. 

[Xilman]

Sure. Give me a little while and I'll post another satellite image and explain how it was taken.

 

[Xilman]

Right, let's see what I can do. First though, some advice on what you can likely do in case you think you need a large aperture telescope which need not be the case. What you do need is a camera. I have a 0.4m aperture reflector, which is larger than average. A 0.2m (8" for those who like imperial units) might be more typical and has 1/4 the collecting area of mine so you will need 4 times the exposure length, all other things (sky brightness, camera sensitivity, etc) being equal. I can get below 20th magnitude in half an hour or so, so most everyone should be able to reach similar faintness within a single night.

First thing to do is dig out an ephemeris of the faint satellites. There are two sites which provide this service. JPL's Horizon's site https://ssd.jpl.nasa.gov/horizons.cgi gives data for all solar system satellites but not, IMO, in as convenient a format as the Minor Planet Center at https://minorplanetcenter.net/iau/NatSats/NaturalSatellites.html not least because the latter can produce ephemerides for several satellites at once; however, it is restricted to the outer satellites of J,S,U and N.

Having found your satellites and checked that they are bright enough and not too close to the Moon, point your telescope there and start taking subs. Unfiltered is generally best to maximise the signal to noise ratio. Keep the subs reasonable short, 30--120s is my choice, for two reasons. First, if you have guiding errors, or an unfortunately placed cosmic ray hit, or are not properly focussed, or are photobombed by an artificial satellite, you haven't wasted much time. Secondly, these objects move across the sky! Too long an exposure and they will be trailed.

Assuming you have a long enough total exposure, where long enough depends on your telescope and the brightness of the satellite, the first step is to examine the subs to throw away any which are unusable for the reasons given above. What I do next is to put a WCS (World Coordinate System) on all the subs FITS files. The best way (IMO) is to use the astrometry.net plate solver. As you will be solving dozens or hundreds of images, it is by far the best to install the plate solver on your own computer.

Next is to stack the images. Most stackers seem to find matching stars in each image and stack on them, which is fine as far as it goes. Unfortunately it doesn't go far enough, as we'll see. Personally I prefer SWarp (https://www.astromatic.net/software/swarp/) because it uses the WCS on each image for alignment purposes.

The first stack will (or should) show circular stars. Load the stack into an image viewer (I use ds9) and push the contrast up as far as seems reasonable. Check the image against the corresponding field of Aladin Lite (https://aladin.u-strasbg.fr/AladinLite/) to ensure that stars at least as faint as the satellite are visible. If not, game over. If the satellite is bright enough, you might see a trail at this stage but you generally will not because the faint light from it is smeared out to invisibility. You need to stack your subs on the predicted motion of the satellite so it appears circular and the stars are trailed. The ephemerides you downloaded earlier will include the predicted motion of the satellite, assuming you asked the web site(s) nicely.

There are two ways of stacking on predicted motion: type in that motion into a stacker which understands such things, or do it yourself. Astrometrica allegedly does it for you but I never had much success. Instead what I do is to run a script which takes a set of FITS files and a sky motion (in arcsec/hr) in RA and DEC of the satellite which then fudges the WCS in each image to move artificially the stars in the opposite direction to and speed of the satellite. That script is available on request. Eventually it will be uploaded to my web site www.astropalma.com but I haven't yet got a round tuit.

The final stage should be obvious: stack on the WCS of the modified FITS files, load into your favourite FITS viewer and/or image processing app. and fiddle with the contrast until you can see the satellite in the correct place. Remember to compare with the Aladin Lite field to ensure that you are not seeing an unfortuinately placed background star or galaxy. Galaxies are surprisingly common below 20th magnitude and start to outnumber stars at about 21-22.

Here is Lysithea, aka Jupiter-X. The twenty 1-minute subs were processed as describe above. Lysithea was relatively bright, at magnitude 18.3, so extreme measures were not required. Note the circular satellite and the stars which are trailed because of the -18.61 arcsec/hour mean motion in RA and -1.63 arcsec/hour in Dec.

[Xilman]

On 16/06/2021 at 16:36, Xilman said:

Some other satellites may be forthcoming, either from last year, or from later this year.

And here is Leda, aka Jupiter-XIII, taken last night. The sky was unusually transparent and the seeing was good so the mag 19.6 satellite shows up very clearly on a stack totaling only 940 seconds. The left side is stacked on the stars and shows a trailed satellite due to its motion. The right was stacked on the predicted motion of Leda and shows a circular satellite and trailed stars.

 

I also tried for Praxidike, or Jupiter-XXVII, spending almost three hours on it.  A serious of gruesome equipment failures (guiding mostly) left insufficient subs for the satellite to show up. It should have been easily visible as it was only 2 magnitudes (6.3 times) fainter.

Edited August 30, 2021 by Xilman
Fix typ

[Xilman]

On 30/08/2021 at 01:02, Xilman said:

I also tried for Praxidike, or Jupiter-XXVII, spending almost three hours on it.  A serious of gruesome equipment failures (guiding mostly) left insufficient subs for the satellite to show up. It should have been easily visible as it was only 2 magnitudes (6.3 times) fainter.

Another attempt was made for Praxidike on 2021-09-06. I had intended to spend 2 hours or so taking data but the TCS died after only 41 1-minute subs had been collected. The resulting stack was just long enough to pick up the satellite but not as clearly as I would have liked.  The image here is the result of stacking on the predicted motion of Jupiter-XXVII. It shows a small circular dot in a field of trailed stars. The measured position is within an arc-second of that predicted. The ephemeris magnitude is V=21.7  which is so faint because the satellite is only 7km across and of low albedo. The variations in the sky background arise from highly contrast-stretched scattered light from Jupiter.

 

 

The camera's cooler has since died and, although it is still usable, the increased dark noise precludes me finding any more faint satellites this season. At least the TCS is working again. Perhaps a better image may be forthcoming next year.